Merge conflict resolution might be time-consuming and lead to defects, compromising development productivity and system quality. Developers might reduce such adverse impacts by avoiding concurrent programming tasks that are more likely to change the same files and cause merge conflicts. As manually predicting such risk is hard, we propose the TAITIr tool, which approximates the set of files changed by a task (task interface) and reports conflict risk whenever there is an intersection between task interfaces. TAITIr uses as input the acceptance tests related to the tasks for predicting file changes, deriving test-based task interfaces. To assess TAITIr’s conflict risk predictions, we measure precision and recall of 6,360 task pairs from 19 Rails projects on GitHub. Our results confirm that the intersection among task interfaces is associated with a higher probability of merge conflict risk. A minimal intersection predicts conflict risk with 0.59 precision and 0.98 recall. We observe that the higher the intersection size, the higher the number of files changed by both tasks. This way, developers might use the intersection size between interfaces as a degree of conflict risk between tasks, choosing a task to work on depending on it. We also find that TAITIr’s predictions outperform predictions based on changed files by similar past tasks. Our analysis derives several other results, considering variations of our notion of an interface in two dimensions: parts of the test code considered for computing interfaces, kinds of files abstracted by the interfaces.
Thu 18 JulDisplayed time zone: Brasilia, Distrito Federal, Brazil change
14:00 - 15:30 | Software Maintenance and Comprehension 3Research Papers / Journal First at Pitomba Chair(s): Xin Xia Huawei Technologies | ||
14:00 18mTalk | Revealing Software Development Work Patterns with PR-Issue Graph Topologies Research Papers Cleidson de Souza Federal University of Pará, Brazil, Emilie Ma University of British Columbia, Jesse Wong University of British Columbia, Dongwook Yoon University of British Columbia, Ivan Beschastnikh University of British Columbia | ||
14:18 18mTalk | Using acceptance tests to predict merge conflict risk Journal First Thaís Rocha UFAPE - Universidade Federal do Agreste de Pernambuco, Paulo Borba Federal University of Pernambuco Pre-print | ||
14:36 18mTalk | Generative AI for Pull Request Descriptions: Adoption, Impact, and Developer Interventions Research Papers Tao Xiao Nara Institute of Science and Technology, Hideaki Hata Shinshu University, Christoph Treude Singapore Management University, Kenichi Matsumoto Nara Institute of Science and Technology Pre-print Media Attached | ||
14:54 18mTalk | SimLLM: Measuring Semantic Similarity in Code Summaries Using a Large Language Model-Based Approach Research Papers | ||
15:12 18mTalk | Sharing Software-Evolution Datasets: Practices, Challenges, and Recommendations Research Papers David Broneske DZHW Hannover, Germany, Sebastian Kittan Otto-von-Guericke Unviersity Magdeburg, Germany, Jacob Krüger Eindhoven University of Technology |